ORACLE_Collect_Collect On-site info for thediagnosing
1. The database
1).Describe the issue 2).Collect the database alert_<ORACLE_SID>.log oracle9i, 10g: # su - oracle $ORACLE_BASE/admin/<ORACLE_SID>/bdump/alert_<ORACLE_SID>.log Or find out the folder: SQL> select value from v$parameter where name = 'user_dump_dest'; oracle11g: $ORACLE_BASE/diag/rdbms/<DB_NAME>/<ORACLE_SID>/trace/alert_<ORACLE_SID>.log If the alert_<ORACLE_SID>.log is so large, please feed back the part of the alert alert_<ORACLE_SID>.log: $ head -20000 alert_<ORACLE_SID>.log > /tmp/alert_<ORACLE_SID>_head.log $ tail -30000 alert_<ORACLE_SID>.log > /tmp/alert_<ORACLE_SID>_tail.log Or find out the folder: SQL>select value from v$diag_info where name = 'Diag Trace'; 3).Collect the listener.log Oracle9i or Oracle 10g: % $ORACLE_HOME/network/log/listener.log oracle11g: % $ORACLE_BASE/diag/tnslsnr/<HOSTNAME>/<LISTENER_NAME>/trace/listener.log % $ORACLE_HOME/network/admin/listener.ora % $ORACLE_HOME/network/admin/sqlnet.ora 4).Collect the pfile % sqlplus '/ as sysdba' SQL>CREATE PFILE=’$ORACLE_HOME/dbs/inittmppfile.ora’ FROM SPFILE; 5).The operation system version #uname -a 6). The database version and patch # su - oracle $ opatch lsinventory -all
|
2. The CRS
1). Check the CRS # su - oracle $ crsctl check crs $ crsctl check cssd $ crsctl check evmd $ crsctl check crsd $ crs_stat –t 2). Check the CRS setup # su - oracle $ cd $ORA_CRS_HOME/bin $ ./cluvfy stage -pre crsinst -n <node1>,<node2> -verbose $ ./cluvfy stage -post crsinst -n <node1>,<node2> -verbose 3). Check the OCR and VOTE # su - oracle $ cd $ORA_CRS_HOME/bin $ ./ocrcheck $ ./ocrconfig -showbackup $ ./crsctl query css votedisk 4). The CRS log (The root user) CRS log: <ORA_CRS_HOME>/log/<hostname> Need to feedback all files excepte the ocssd history log: <ORA_CRS_HOME>/log/<hostname>/cssd <====== Just feedback some files in the recent days. oprocd log: # cd /etc/oracle/oprocd # ls -lt|grep oprocd <====== Feedback some files in the recent days.
|
3. The cluster log
1). Linux VCS # tar -cvf vcs_log.tar /var/VRTSvcs/log 2). SUN CLUSTER /var/adm/messages /var/opt/SUNWscor/oracle_listener/*log* /var/opt/SUNWscor/oracle_server/*log* 3). HP CLUSTER /etc/cmcluster/cmclconf.ascii /etc/cmcluster/PKG_NAME/pkg1conf.asci /etc/cmcluster/PKG_NAME/control.sh /etc/cmcluster/PKG_NAME/control.sh.log /etc/cmcluster/PKG_NAME/spy.sh.log /etc/cmcluster/PKG_NAME/spy.sh.log /var/adm/cmcluster/fr.dump.cmld.x <===== ‘ x’ The max one # cmviewcl -v 4). AIX HACMP /var/hacmp/log/hacmp.out or /tmp/hacmp.out /var/hacmp/adm/cluster.log or find . -name cluster.log /usr/es/sbin/cluster/history/cluster.mmddyyyy /var/hacmp/log/clstrmgr.debug.long /var/hacmp/log/cspoc.log.long
|
4. The operation system log
1). Linux /var/log/messages 2). Sun Solaris /var/adm/messages 3). HP-UX /var/adm/syslog/syslog.log Tru64: /var/adm/messages 4). IBM AIX /bin/errpt -a > messages.out
|
5. The operation systemparameters
1). Linux SUSE # sysctl -a # cat /etc/sysctl.conf # ulimit -a # cat /etc/security/limits.conf 2). SUN solaris # cat /etc/system 3). HP-UX # sam And ‘ENTER’. 4). IBM AIX $ lsattr -E -l sys0 -a maxuproc $ env |grep -i scope # lsattr -E -l sys0|grep out # vmo -aF
|
6. The AWR/ADDM/ASH report
How to collect the AWR_ADDM_ASH reports.docx
7. The current OS resourcesinfo(Come from the OSW)
Linux) # iostat -xk 1 3 # vmstat 1 3 # eval top -b -n 1 | head -50 # ps -elf # mpstat -P ALL 1 2 # cat /proc/meminfo # cat /proc/slabinfo HP-UX|HI-UX) # iostat 1 3 # vmstat 1 3 # top -d 1 # ps -elf # sar -A -S 1 2 SunOS) # iostat -xn 1 3 # vmstat 1 3 # prstat 1 2 # ps -elf # mpstat 1 2 AIX) # iostat -D 1 3 # vmstat 1 3 # top -Count 1 # ps -elf # mpstat 1 2 OSF1) # iostat 1 3 # vmstat 1 3 # top -d1 # ps -elf # sar -S |
8. The OSW info
# cd /opt # mkdir OSWbb # chmod 777 -R OSWbb <====== The folder size: 1.5G~2.0G/One day Upload the oswbb5.tar to /opt/OSWbb # cd OSWbb # tar xvf oswbb5.tar # cd oswbb # vi private.net
# chown oracle:oinstall private.net # chmod 777 private.net
If you want to collect the OSW information: # nohup ./startOSWbb.sh 3 10 & <====== collect data at 3 second intervals, and log the last 10 hours of data to archive files
If you want to stop OSW, you can to do this: ./stopOSWbb.sh
You can find out the OSW log in the /opt/OSWbb/oswbb/archive |